The energy consumption of the bitcoin network is not inherent to proof of work as a consensus mechanism. Bitcoin uses a difficulty setting to keep the rate of block creation relatively constant even as the hashing power dedicated to it increases. The reason so much energy is being used boils down to speculation in the price of bitcoin. At $10,000 per coin and with a block reward of 12.5 coins, successfully mining a block yields $125,000. That means that for each individual miner it would be profitable to spend up to that much per block they win. $125k can buy a LOT of electricity. But on the other hand, if the price of bitcoin were closer to $1 again and nobody expected it to appreciate very much then the level of investment, and thereby power usage, would go way down.